Add perf metrics for 2.43.0

[khluu]

REGRESSION 47.69%: tasks_per_second (THROUGHPUT) regresses from 427.6684054263659 to 223.72283336753296 in benchmarks/many_nodes.json
REGRESSION 39.45%: pgs_per_second (THROUGHPUT) regresses from 22.74259799982883 to 13.769920950065924 in benchmarks/many_pgs.json
REGRESSION 32.47%: tasks_per_second (THROUGHPUT) regresses from 588.1874987887217 to 397.2166130464684 in benchmarks/many_tasks.json
REGRESSION 28.32%: multi_client_put_gigabytes (THROUGHPUT) regresses from 47.38817873082456 to 33.968763685249776 in microbenchmark.json
REGRESSION 24.93%: single_client_tasks_and_get_batch (THROUGHPUT) regresses from 7.56687497857023 to 5.680486486924976 in microbenchmark.json
REGRESSION 15.78%: single_client_wait_1k_refs (THROUGHPUT) regresses from 5.424179804481537 to 4.568303345800756 in microbenchmark.json
REGRESSION 12.61%: single_client_get_object_containing_10k_refs (THROUGHPUT) regresses from 12.987017792446045 to 11.349304359077122 in microbenchmark.json
REGRESSION 10.56%: 1_1_async_actor_calls_async (THROUGHPUT) regresses from 4669.450952977499 to 4176.255856784519 in microbenchmark.json
REGRESSION 9.48%: n_n_actor_calls_with_arg_async (THROUGHPUT) regresses from 2732.074477927061 to 2473.163327924083 in microbenchmark.json
REGRESSION 8.44%: single_client_tasks_sync (THROUGHPUT) regresses from 1013.1673399687909 to 927.6195505025511 in microbenchmark.json
REGRESSION 8.24%: 1_1_async_actor_calls_with_args_async (THROUGHPUT) regresses from 2953.9546990180343 to 2710.6889492331006 in microbenchmark.json
REGRESSION 6.69%: single_client_tasks_async (THROUGHPUT) regresses from 8032.409007811969 to 7495.427972759755 in microbenchmark.json
REGRESSION 5.90%: multi_client_tasks_async (THROUGHPUT) regresses from 22745.167201851888 to 21404.22305204687 in microbenchmark.json
REGRESSION 3.88%: 1_1_async_actor_calls_sync (THROUGHPUT) regresses from 1474.7069154202795 to 1417.4714984406974 in microbenchmark.json
REGRESSION 3.69%: 1_1_actor_calls_sync (THROUGHPUT) regresses from 1985.8445748508243 to 1912.6174326068551 in microbenchmark.json
REGRESSION 3.63%: n_n_async_actor_calls_async (THROUGHPUT) regresses from 23390.23156817461 to 22542.16458297063 in microbenchmark.json
REGRESSION 3.53%: 1_1_actor_calls_concurrent (THROUGHPUT) regresses from 5218.943213086157 to 5034.662033195711 in microbenchmark.json
REGRESSION 3.40%: client__1_1_actor_calls_sync (THROUGHPUT) regresses from 522.5361491166506 to 504.79511779286827 in microbenchmark.json
REGRESSION 2.70%: client__tasks_and_put_batch (THROUGHPUT) regresses from 14255.363968554979 to 13870.502654856362 in microbenchmark.json
REGRESSION 2.43%: multi_client_put_calls_Plasma_Store (THROUGHPUT) regresses from 15931.811977493457 to 15543.88478060201 in microbenchmark.json
REGRESSION 1.77%: single_client_put_calls_Plasma_Store (THROUGHPUT) regresses from 4866.041059585032 to 4780.133410691611 in microbenchmark.json
REGRESSION 1.58%: client__put_gigabytes (THROUGHPUT) regresses from 0.15408690555366225 to 0.1516460375456968 in microbenchmark.json
REGRESSION 0.87%: n_n_actor_calls_async (THROUGHPUT) regresses from 26441.672940245888 to 26212.485337497925 in microbenchmark.json
REGRESSION 0.22%: single_client_get_calls_Plasma_Store (THROUGHPUT) regresses from 10611.609624248378 to 10588.032610459411 in microbenchmark.json
REGRESSION 190.22%: stage_4_spread (LATENCY) regresses from 0.20731175121781154 to 0.6016603078317008 in stress_tests/stress_test_many_tasks.json
REGRESSION 64.03%: avg_iteration_time (LATENCY) regresses from 0.7432218527793885 to 1.2191412186622619 in stress_tests/stress_test_dead_actors.json
REGRESSION 55.33%: stage_0_time (LATENCY) regresses from 4.678402662277222 to 7.266809701919556 in stress_tests/stress_test_many_tasks.json
REGRESSION 52.14%: stage_3_creation_time (LATENCY) regresses from 1.1666333675384521 to 1.7748794555664062 in stress_tests/stress_test_many_tasks.json
REGRESSION 29.29%: dashboard_p99_latency_ms (LATENCY) regresses from 3492.239 to 4515.016 in benchmarks/many_actors.json
REGRESSION 16.55%: avg_pg_remove_time_ms (LATENCY) regresses from 1.2978452747749973 to 1.5126715450448436 in stress_tests/stress_test_placement_group.json
REGRESSION 16.10%: dashboard_p50_latency_ms (LATENCY) regresses from 4.297 to 4.989 in benchmarks/many_nodes.json
REGRESSION 8.46%: 10000_args_time (LATENCY) regresses from 17.283816822999995 to 18.745566793000002 in scalability/single_node.json
REGRESSION 4.86%: dashboard_p95_latency_ms (LATENCY) regresses from 2727.324 to 2859.782 in benchmarks/many_actors.json
REGRESSION 4.18%: 10000_get_time (LATENCY) regresses from 23.877336194999998 to 24.874795295 in scalability/single_node.json
REGRESSION 3.53%: avg_pg_create_time_ms (LATENCY) regresses from 1.470244198198109 to 1.522115139639421 in stress_tests/stress_test_placement_group.json
REGRESSION 3.36%: 1000000_queued_time (LATENCY) regresses from 192.979472547 to 199.46889562500002 in scalability/single_node.json
REGRESSION 2.39%: stage_3_time (LATENCY) regresses from 1823.5409452915192 to 1867.0339968204498 in stress_tests/stress_test_many_tasks.json
REGRESSION 0.27%: stage_1_avg_iteration_time (LATENCY) regresses from 12.529016280174256 to 12.563214778900146 in stress_tests/stress_test_many_tasks.json

[jjyao]

REGRESSION 47.69%: tasks_per_second (THROUGHPUT) regresses from 427.6684054263659 to 223.72283336753296 in benchmarks/many_nodes.json

Expected due to release test runtime change.

REGRESSION 39.45%: pgs_per_second (THROUGHPUT) regresses from 22.74259799982883 to 13.769920950065924 in benchmarks/many_pgs.json

Expected due to release test runtime change.

REGRESSION 32.47%: tasks_per_second (THROUGHPUT) regresses from 588.1874987887217 to 397.2166130464684 in benchmarks/many_tasks.json

Expected due to release test runtime change.

REGRESSION 28.32%: multi_client_put_gigabytes (THROUGHPUT) regresses from 47.38817873082456 to 33.968763685249776 in microbenchmark.json

Expected due to release test runtime change.

REGRESSION 24.93%: single_client_tasks_and_get_batch (THROUGHPUT) regresses from 7.56687497857023 to 5.680486486924976 in microbenchmark.json

Expected due to release test runtime change.

REGRESSION 15.78%: single_client_wait_1k_refs (THROUGHPUT) regresses from 5.424179804481537 to 4.568303345800756 in microbenchmark.json

Noise + small regression that doesn't block the release

REGRESSION 12.61%: single_client_get_object_containing_10k_refs (THROUGHPUT) regresses from 12.987017792446045 to 11.349304359077122 in microbenchmark.json

Noise

REGRESSION 10.56%: 1_1_async_actor_calls_async (THROUGHPUT) regresses from 4669.450952977499 to 4176.255856784519 in microbenchmark.json

Noise

REGRESSION 190.22%: stage_4_spread (LATENCY) regresses from 0.20731175121781154 to 0.6016603078317008 in stress_tests/stress_test_many_tasks.json

Expected due to release test runtime change.

REGRESSION 64.03%: avg_iteration_time (LATENCY) regresses from 0.7432218527793885 to 1.2191412186622619 in stress_tests/stress_test_dead_actors.json

REGRESSION 55.33%: stage_0_time (LATENCY) regresses from 4.678402662277222 to 7.266809701919556 in stress_tests/stress_test_many_tasks.json

Expected due to release test runtime change.

REGRESSION 52.14%: stage_3_creation_time (LATENCY) regresses from 1.1666333675384521 to 1.7748794555664062 in stress_tests/stress_test_many_tasks.json

Expected due to release test runtime change.

REGRESSION 16.55%: avg_pg_remove_time_ms (LATENCY) regresses from 1.2978452747749973 to 1.5126715450448436 in stress_tests/stress_test_placement_group.json

Noise.

[aslonnie]

will prefer #50894 which is more up to date.

numbers look similar (at least to me)